Note that, during the recent reddit AMA, Elon announced that the current (post-Adelaide) BFS design now has a third center engine with an expansion ratio between the sea-level and vacuum versions.paperburn1 wrote:But remember it has to fill 4 different jobs with the engine selection. Sub-orbital point to point, orbital point to point, LEO at least to space station range with 173 ton payload and MARS.
At least according to grand plan using only two different types of engine.
After running a few of these scenarios, I'm a lot less convinced that there's only going to be one version of the BFS. We know (also from the AMA) that they're planning on changing the payload structure substantially for the tanker versions, and the long-haul crewed version will also obviously be substantially different from the cargo version. If they're going to fool with payload structure, why not fool with the engines as well?
Here's the other use case that isn't covered with the current engining for the BFS: landing fail-safe on the Moon. Suppose you want to do a powered descent with only 10 tonnes of payload and enough fuel to return to Earth:
Delta-v from lunar surface to TEI (plus 500 m/s for landing): 3360 m/s
Assume dry mass of 85 t and no payload on return flight, and you get a mass ratio of exp(3360/(9.8*375)) = 2.49, which means you need 127 t of propellant. That makes mass to the lunar surface 85t + 127t + 10t = 222t.
For safe powered descent, you need to be able to hover, which means that thrust = lunar weight. Lunar weight = 222t * 1.62 m/s^2 = 360 kN. But a single SL Raptor running in vacuum is going to generate about 1830 kN (this is a bit of a SWAG, but it's definitely more than 1700 and less than 1900), so you can only land with a single engine throttled to 20%, which is right at the very bottom of the stated design range for the Raptor.
And you simply can't do it if you're adhering to the "two-engine fail-safe" criterion that Elon's been touting. To get to that, you have to have 444t of wet mass to the surface, which would be 232t of payload, which is more payload than you can launch to LEO.
My guess is that, if they really want to go after all of these markets (Earth orbital, Earth point-to-point, Martian interplanetary, lunar), they're going to need at least three variants:
1) The current plan of record: 4 Raptor vac, 2 Raptor SL, and the 1 Raptor intermediate that Musk talked about in the reddit AMA. This'll handle Earth orbital (both up and down) and Martian.
2) A single-stage-to-suborbit version with Raptor intermediate engines for Earth point-to-point, to make something safe enough to pass common carrier regulatory standards. Per the analysis up above, this is right on the hairy edge of feasible, and may require some additional structural compromises.
3) A lunar version, similar to #1, but with the three center engines replaced by a sub-scale Raptor that only generates about 800 kN of max thrust. That'll let it land light payloads on the Moon but still also handle a small amount of lunar down-mass (a crew, for instance) to land fail-safe on Earth.
If they can figure out how to do the SSTSO on only seven intermediate engines, they could use a single common thrust structure and just mix-and-match engines. Again, that needs a lot of analysis to figure if it's possible, but this whole idea of launching passengers on a regular basis with a booster that contains 31 fate-shared rocket engines and no launch escape is just silly. It's bad enough with only seven engines, but that's at least on the outer edge of sane.